Comparing comorbidity scales: Attending physician score versus the Cumulative Illness Rating Scale for Geriatrics

Assessing comorbidity using the Cumulative Illness Rating Scale for Geriatrics (CIRS-G) and its comprehensive manual is time consuming. We investigated if similar information could be obtained by a simpler assessment based on the original CIRS.

Data from a randomized chemotherapy trial (RCT) on advanced NSCLC (non-small cell lung cancer) were analyzed. Baseline comorbidity was assessed by 1) trained oncologists using hospital records and the CIRS-G manual (CIRS-G), 2) by patients' oncologists/pulmonologists (local investigators=LI-score) using a brief set of instructions. By both methods, the severity of comorbidity in 14 organ systems was graded 0 (no problem) to 4 (extremely severe). The agreement between methods was assessed using Bland-Altman analysis and weighted kappa statistics. The impact of comorbidity on survival was analyzed by Cox regression.

Complete data were available for 375/446 (84%) patients enrolled in the RCT. Median age was 65years (25-85). Overall, more comorbidities and higher severity were registered by the CIRS-G compared to the LI-score. Severe comorbidity was registered for 184 (49%) and 94 (25%) patients according to the CIRS-G and LI-scores, respectively. Mean total score was 7.0 (0-17) (CIRS-G) versus 4.2 (0-16) (LI-score), and mean severity index (total score/number of categories with score >0) was 1.73 (SD 0.46) versus 1.43 (SD 0.78). Neither the CIRS-G scores nor the LI-scores were prognostic for survival.

The CIRS-G scores and LI-scores had poor agreement, indicating that assessment method affects the registration of comorbidity. Thorough descriptions of comorbidity registrations in trials are paramount due to lack of a standardized assessment.